DC power supply intelligent illumination control method and system

Abstract

The invention relates to the technical field of illumination intelligent control, in particular to a direct-current power supply intelligent illumination control method and system, and the method comprises the steps: carrying out the human body existence detection of a target region, and adjusting the illumination of the target region to a first brightness if a detection result represents that a human body exists in the target region, and maintaining a preset duration in the state of the first brightness, performing stay behavior detection on the target area, if a behavior detection result represents that a person in the target area has a stay behavior, adjusting the illumination of the target area to a second brightness, and if the illumination of the target area is in the state of the second brightness, adjusting the illumination of the target area to a second brightness. If the existence detection result shows that no human body exists in the target area, lighting of the target area is turned off. During illumination control, passing and staying scenes are accurately distinguished through double-stage logic of human body existence detection and staying behavior detection, and the problem that in the prior art, a control method of a direct-current power supply illumination system is not intelligent enough is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent lighting control, and particularly to a DC-powered intelligent lighting control method and system. Background Art

[0002] In modern lighting systems, the DC power supply method is gradually showing many significant advantages. For example, DC power supply can provide more stable power support for lighting devices, with low energy loss and no need for complex AC-DC conversion equipment. The most prominent advantage of DC power supply is that it can achieve more precise intelligent control and dimming performance.

[0003] However, the current lighting systems have obvious limitations in brightness adjustment. Most current mainstream products adopt the basic logic of "turn on the light when a person is detected and turn off the light after a certain period of time". This mode is similar to a set mechanical switch - it can only perform fixed actions and cannot handle complex situations. There are often two typical problems in actual applications: when someone passes by quickly, the light stays on continuously due to the detection of the movement signal, resulting in waste; and when a person stays still, the light may turn off prematurely due to the disappearance of the movement signal. This reflects that the current system lacks the ability to deeply understand the behavior pattern and is difficult to make reasonable judgments according to the actual usage scenario.

[0004] Therefore, people need a more intelligent DC-powered intelligent lighting control method. Summary of the Invention

[0005] Therefore, the present invention provides a DC-powered intelligent lighting control method and system to solve the problem that the control method of the DC-powered lighting system in the prior art is not intelligent enough.

[0006] The present invention provides a DC-powered intelligent lighting control method, including: Performing human presence detection on a target area to obtain a presence detection result; If the presence detection result indicates that there is a human in the target area, adjust the lighting of the target area to a first brightness and maintain it for a preset duration in the state of the first brightness; If the presence detection result indicates that there is a human in the target area, perform a stay behavior detection on the target area to obtain a behavior detection result, wherein the operation duration of the stay behavior detection is less than the preset duration; If the behavior detection result indicates that the person in the target area has a stay behavior, adjust the lighting of the target area to a second brightness, wherein the second brightness is higher than the first brightness; If the lighting of the target area is in the state of the second brightness and the presence detection result indicates that there is no human in the target area, turn off the lighting of the target area.

[0007] In a preferred solution: perform human presence detection on the target area to obtain a presence detection result, including: Obtain a human presence sensing signal, where the human presence sensing signal includes at least one of infrared data, millimeter-wave radar data, image data, ultrasonic data, pressure data, or vibration data; Obtain a presence detection result according to the human presence sensing signal.

[0008] In a preferred solution: perform a stay behavior detection on the target area to obtain a behavior detection result, including: Obtain the position status information of the people in the target area; Obtain a behavior detection result according to the position status information of the people.

[0009] In a preferred solution: obtain a behavior detection result according to the position status information of the people, including: Obtain the environmental information of the target area; Based on a preset neural network model, obtain a behavior detection result according to the position status information of the people and the environmental information.

[0010] In a preferred solution: based on a preset neural network model, obtain a behavior detection result according to the position status information of the people and the environmental information, including: Obtain the direction angle and velocity component of the people relative to the exit of the target area according to the people status information and the environmental information; Establish input data according to the direction angle and velocity component; Input the input data into the preset neural network model to obtain a behavior detection result.

[0011] In a preferred solution: the preset neural network model includes multiple sub-neural network modules and a comprehensive decision-making module, where the multiple sub-neural network modules correspond one by one to multiple exits in the target area and are respectively used to input the input data corresponding to each exit, and the comprehensive decision-making module is connected to the output ends of the multiple sub-neural network modules and is used to output a behavior detection result.

[0012] In a preferred solution: the sub-neural network module includes an LSTM layer, the sub-neural network is used to output a first context vector, the comprehensive decision-making module includes a feedforward neural network layer, and the comprehensive decision-making module is used to input a second context vector formed by weighted summation of multiple first context vectors.

[0013] The present invention also provides a DC-powered intelligent lighting control system, including: A presence detection module, configured to perform human presence detection on the target area to obtain a presence detection result; The first adjustment module is used to adjust the illumination of the target area to the first brightness when the detection result indicates the presence of a human body in the target area, and maintain the first brightness state for a preset duration; The behavior detection module is used to perform a stay behavior detection on the target area when the detection result indicates the presence of a human body in the target area, and obtain a behavior detection result. Among them, the operation duration of the stay behavior detection is less than the preset duration; The second adjustment module is used to adjust the illumination of the target area to the second brightness when the behavior detection result indicates that the person in the target area has a stay behavior, where the second brightness is higher than the first brightness; The third adjustment module is used to turn off the illumination of the target area when the illumination of the target area is in the second brightness state and the detection result indicates the absence of a human body in the target area.

[0014] In a preferred solution: it further includes: The timing control module is used to control the illumination of the target area according to time periods; The remote control module is used to remotely adjust and view the illumination state of the target area.

[0015] The present invention also provides a computer-readable storage medium for storing computer-readable programs or instructions. When the programs or instructions are executed by a processor, the steps in any one of the above DC-powered intelligent lighting control methods can be implemented.

[0016] The beneficial effects of adopting the above embodiments are: The present invention provides a DC-powered intelligent lighting control method and system. It performs a human presence detection on the target area to obtain a presence detection result. If the presence detection result indicates the presence of a human body in the target area, the illumination of the target area is adjusted to the first brightness and maintained in the first brightness state for a preset duration. And a stay behavior detection is performed on the target area to obtain a behavior detection result. If the behavior detection result indicates that the person in the target area has a stay behavior, the illumination of the target area is adjusted to the second brightness, where the second brightness is higher than the first brightness. If the illumination of the target area is in the second brightness state and the presence detection result indicates the absence of a human body in the target area, the illumination of the target area is turned off. When controlling the illumination, the present invention uses a two-stage logic of "human presence detection + stay behavior detection" to accurately distinguish between passing-by and staying scenarios, realizing dynamic dimming for maintaining the basic brightness and improving the comfortable brightness, solving defects such as misjudgment and energy consumption waste in traditional solutions, and solving the problem that the control method of the DC-powered lighting system in the prior art is not intelligent enough. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a method flow chart of the DC-powered intelligent lighting control method in the present invention; Figure 2 The specific process of human presence detection in the present invention; Figure 3 The specific structure of the preset neural network model in the present invention; Figure 4 The system structure diagram of the DC-powered intelligent lighting control system in the present invention. Detailed implementation manners

[0018] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0019] In conjunction with Figure 1 As shown, a specific embodiment of the present invention discloses a DC-powered intelligent lighting control method, including: S101. Perform human presence detection on the target area to obtain a presence detection result; S102. If the presence detection result indicates that there is a human in the target area, adjust the lighting of the target area to the first brightness and maintain it for a preset duration in the state of the first brightness; S103. If the presence detection result indicates that there is a human in the target area, perform a stay behavior detection on the target area to obtain a behavior detection result, where the operation duration of the stay behavior detection is less than the preset duration; S104. If the behavior detection result indicates that the person in the target area has a stay behavior, adjust the lighting of the target area to the second brightness, where the second brightness is higher than the first brightness; S105. If the lighting of the target area is in the state of the second brightness and the presence detection result indicates that there is no human in the target area, turn off the lighting of the target area.

[0020] During the above process, human presence detection is only used to detect whether there is a human body in the target area, while stay behavior detection is used to detect the specific behavior of people in the target area, mainly to determine whether it is a passing behavior or a stay behavior, and it needs to more specifically judge the intention of people. It can be understood that the method process of human presence detection is simpler than that of stay behavior detection. The brightness represented by the first brightness is lower than the second brightness. The first brightness only needs to provide basic visibility, while the second brightness is the optimal brightness required when people perform operations and rest indoors. It can be understood that the first brightness and the second brightness can be determined according to actual ambient light, time and other factors. The specific adjustment method can be flexibly designed according to different actual situations (such as hardware conditions, cost considerations, software strategies, etc.). The specific method is the prior art that can be understood by those skilled in the art, so it will not be described in detail in this article.

[0021] For the method provided in the above process, when it is detected that there is a human body in the target area, the lighting system can be quickly adjusted to the first brightness. At this time, the first brightness is mainly used to help people distinguish the basic direction when people pass by. For example, when people get up at night and pass through the living room from the bedroom to the bathroom. After detecting the human body of the target area type, the stay behavior detection starts. If it is detected that there is no stay behavior, at this time, the target area will be closed after maintaining the first brightness for a period of time (i.e., the preset duration) to achieve timely closing after people leave and avoid the phenomenon of "constant lighting". Only after detecting that there is a stay behavior, the lighting system is adjusted to the second brightness and maintained continuously.

[0022] It can be understood that because the human presence detection is simple to implement, it can run continuously. Only when the lighting in the target area is in the state of the second brightness and the detection result indicates that there is no human body in the target area, the lighting will be turned off, avoiding the situation where the lights are automatically turned off when people are stationary (such as working, reading).

[0023] Generally speaking, this embodiment first confirms whether there is a human body in the area, and then further judges whether it is a stay behavior, avoiding misjudgment caused by a single movement signal in the traditional scheme (such as passing by quickly or pet activities). It can distinguish between "brief passing by" and "continuous staying", solving the defect of the "one-size-fits-all" dimming logic of the existing system. For example, in the passing scenario, only the lowest brightness (the first brightness) is maintained to reduce energy waste. In the staying scenario, it can automatically smoothly transition from the first brightness to the second brightness, avoiding visual discomfort caused by sudden brightness changes in the traditional scheme and meeting the actual use requirements. The logic loop of this embodiment can avoid abnormal working conditions, and the setting of the preset duration can achieve safety backup, that is, the maintenance duration is set in the first brightness stage. Even if the stay detection module fails briefly, it can exit the low-efficiency lighting state through the time threshold to ensure the robustness of the system.

[0024] Specifically, in the above step S101, human presence detection is performed on the target area to obtain a presence detection result, including: Obtain a human presence sensing signal, where the human presence sensing signal includes at least one of infrared data, millimeter-wave radar data, image data, ultrasonic data, pressure data, or vibration data; Based on the human presence sensing signal, obtain a presence detection result.

[0025] In this embodiment, in the human presence detection link, by fusing multi-modal sensing data such as infrared, millimeter-wave radar, image, ultrasonic, pressure, and vibration, the reliability and adaptability of the detection are significantly improved: the complementary multi-source data effectively reduces the false alarm rate of a single sensor, and the millimeter-wave radar and image data enhance the dynamic response and precise positioning capabilities, adapting to the requirements of complex environments and diverse scenarios, and having a good anti-interference design (such as millimeter waves penetrating walls and local image processing). It can be understood that since the human presence detection in this embodiment is only used to detect whether a human exists, complex algorithms are not required, which can effectively reduce the cost of the entire lighting system.

[0026] Further, in a preferred embodiment, in the above step S103, a stay behavior detection is performed on the target area to obtain a behavior detection result, specifically including: Obtain the position status information of the person in the target area; Based on the position status information of the person, obtain a behavior detection result.

[0027] In the above process, the position status information of the person is used to represent the spatio-temporal information such as the position and speed of the person in the target area. A more specific and easy-to-implement example of this embodiment is: detecting the walking direction and speed of a person in the target area. If the direction remains unchanged and the speed is relatively fast, it can be roughly considered that the person is just passing by and does not intend to stay in the target area.

[0028] However, obviously the accuracy of the above process is general. Therefore, the present invention also provides a more accurate stay behavior detection method. In a preferred embodiment, the above step: based on the position status information of the person, obtain a behavior detection result, specifically including: Obtain the environmental information of the target area; Based on the position status information of the person and the environmental information, and based on a preset neural network model, obtain a behavior detection result.

[0029] In this embodiment, the introduction of environmental information makes up for the deficiency of simple spatio-temporal feature analysis, and then uses the neural network model to learn the correlation between complex behavior patterns and environmental features, such as distinguishing between standing still and reading and queuing, and identifying instantaneous speed changes caused by collisions, etc., to achieve more refined behavior classification and improve the recognition accuracy of user behavior patterns.

[0030] A more specific, simple and feasible example of this embodiment is as follows: using the environmental image as the environmental information, using image recognition algorithms such as YOLO as the preset neural network model, identifying the areas where people may stay (such as sofas, dining tables) and the areas where people may leave (such as the door) in the environment through the preset neural network model, and then judging whether there is a staying intention according to the spatial relationship between the person position status information and these staying areas.

[0031] However, it can be understood that the above content needs to utilize the collected environmental image, and in the lighting scene control, collecting images may involve user privacy issues. Therefore, to avoid such problems, this embodiment also provides a more practical solution.

[0032] Specifically, as Figure 2 shown, the above steps: obtaining the behavior detection result based on the preset neural network model according to the person position status information and the environmental information specifically include: S201. Obtain the direction angle and velocity component of the person relative to the exit of the target area according to the person status information and the environmental information; S202. Establish the input data according to the direction angle and the velocity component; S203. Input the input data into the preset neural network model to obtain the behavior detection result.

[0033] In the above process, the direction angle refers to the angle between the current traveling direction of the person and the line connecting the current position of the person and the exit of the target area.

[0034] Since the lighting system is mostly installed indoors, and when a person is performing a passing behavior (such as passing through a corridor), the exits for leaving the target area are often limited. Moreover, when a person passes through an area, they will show obvious purposefulness. Therefore, the walking direction and speed of a person when passing through an area will also show characteristics different from other behaviors. This embodiment utilizes this characteristic to detect the staying behavior.

[0035] For example, generally speaking, when a person walks from one exit to another, due to the clear purpose of "passing through", the walking direction will be towards the exit, and the consistency and stability of the data such as the walking speed and direction will be higher than the data in other states such as wandering and doing cleaning.

[0036] In addition, more importantly, the direction angle and velocity component of the person relative to the exit of the target area can be obtained by using non-image recognition methods such as setting ultrasonic sensors at the entrances and exits, avoiding privacy issues.

[0037] More specifically, in combination with Figure 3As shown, in a preferred utility model, the above-mentioned preset neural network model includes multiple sub-neural network modules and a comprehensive decision-making module. Among them, the multiple sub-neural network modules correspond one-to-one with multiple exits in the target area, and are respectively used to input the input data corresponding to each exit. The comprehensive decision-making module is connected to the output ends of the multiple sub-neural network modules and is used to output the behavior detection result.

[0038] In the above process, corresponding input data is established for each exit respectively, and there is a one-to-one correspondence among the input data, the sub-neural network module, and the environmental exit.

[0039] It can be understood that in practice, there are often multiple entrances and exits in the target area. Therefore, in this embodiment, the preset neural network model is further split into multiple sub-neural network modules and a comprehensive decision-making module. Among them, each sub-neural network module is only used to judge whether there is a passing behavior of the person in the target area relative to an exit, and the comprehensive decision-making module is used to summarize the judgment results of multiple sub-neural networks for comprehensive decision-making. The comprehensive decision-making module can be implemented by simple conditional judgment. For example, when all sub-neural network modules believe that there is no passing behavior, the comprehensive decision-making module can determine that the person has a staying intention. On the contrary, if at least one sub-neural network module determines that there is a passing behavior, the comprehensive decision-making module can determine that the person has a passing intention.

[0040] This embodiment further optimizes the design and demonstrates significant advantages in complex multi-exit scenarios through the cooperation of multiple sub-neural network modules and the comprehensive decision-making module: on the one hand, each sub-neural network module focuses on the input data of a single exit (such as the direction angle, speed component, etc.), and can more accurately capture the relative position and movement intention of the person and a specific exit, avoiding misjudgment caused by global feature extraction in multi-exit scenarios; on the other hand, through the summary and conditional judgment of the comprehensive decision-making module, combined with the output results of all sub-network modules, the limitations of a single perspective (such as local occlusion or instantaneous speed change) can be effectively eliminated, and a globally optimal behavior judgment can be achieved. In addition, multiple sub-network modules can process their respective tasks in parallel, significantly reducing the calculation delay, meeting the real-time requirements. When adding a new exit, only the corresponding sub-network module needs to be added, without reconstructing the overall model, and the adaptability is stronger. Each sub-network module operates independently, and a failure will not affect the global decision-making. At the same time, it supports the independent optimization of the sub-network of a specific exit, improving the system robustness. Overall, this solution realizes the balance of high precision, high efficiency, and high robustness in the judgment of staying behavior in multi-exit scenarios through the combination of modularization and conditional judgment, providing a reliable solution for the intelligent perception of complex spaces.

[0041] It is understandable that the specific implementation manners of the above content, such as how the neural network model is preset, how multiple sub-neural network modules make judgments, and how the comprehensive decision-making module conducts comprehensive analysis, are all prior arts that can be understood by those skilled in the art. Therefore, no further description will be given herein.

[0042] The present invention further provides a more specific embodiment. Among them, the sub-neural network module includes an LSTM layer, and the sub-neural network is used to output a first context vector. The comprehensive decision-making module includes a feed-forward neural network layer, and the comprehensive decision-making module is used to input a second context vector formed by weighted summation of multiple first context vectors.

[0043] This embodiment further improves the detection accuracy and decision-making ability of the solution by introducing the deep learning architecture of the LSTM layer and the feed-forward neural network layer. Specifically, the LSTM layer can effectively capture the temporal features of the input data in the sub-neural network module (such as the time series of the speed and position changes of a person), extract long-term and short-term dependence information (such as the standing still, turning back or wandering behavior of a person), so as to generate a first context vector with high semantic expression ability, accurately reflecting the behavior intention of the person and a single exit; the feed-forward neural network layer constructs a second context vector with a global perspective by synthesizing multiple first context vectors after weighted summation (which can be designed to use the statistical frequency of users leaving each exit as the weight), significantly enhancing the association analysis ability of the comprehensive decision-making module for multi-exit scenarios, and being able to effectively distinguish complex behavior patterns (such as a person repeatedly moving or standing still between multiple exits and then choosing an exit). In addition, this solution reduces the misjudgment rate caused by instantaneous speed fluctuations or position jumps through the temporal modeling ability of the LSTM layer; realizes the dynamic adjustment of the priorities of different exits (such as the exit closer to the crowded area has a higher weight) through the weight allocation mechanism of the feed-forward neural network layer, optimizing the flexibility and adaptability of decision-making; under the distributed computing architecture, the independent calculations of the LSTM layer and the feed-forward neural network layer further improve the inference efficiency, meeting the real-time requirements. Overall, this solution realizes high-precision, high-fault-tolerant and high-expandable behavior detection in complex multi-exit scenarios through the combination of temporal modeling, global decision-making and dynamic weights, providing reliable technical support for intelligent space management.

[0044] It is understandable that the principles of the LSTM technology and the feed-forward neural network layer are prior arts that can be understood by those skilled in the art. Therefore, no further description will be given herein.

[0045] Combined with Figure 4 as shown, the present invention further provides a DC-powered intelligent lighting control system, including: A presence detection module 410, configured to perform human presence detection on a target area to obtain a presence detection result; The first adjustment module 420 is configured to adjust the illumination of the target area to a first brightness when the presence detection result indicates that there is a human body in the target area, and maintain the first brightness for a preset duration; The behavior detection module 430 is configured to perform a stay behavior detection on the target area when the presence detection result indicates that there is a human body in the target area, and obtain a behavior detection result, wherein the operation duration of the stay behavior detection is less than the preset duration; The second adjustment module 440 is configured to adjust the illumination of the target area to a second brightness when the behavior detection result indicates that the person in the target area has a stay behavior, wherein the second brightness is higher than the first brightness; The third adjustment module 450 is configured to turn off the illumination of the target area when the illumination of the target area is in the state of the second brightness and the presence detection result indicates that there is no human body in the target area.

[0046] It should be noted here that: the corresponding system provided in the above embodiment can implement the technical solutions described in the above method embodiments. The specific implementation principles of the above modules or units can be referred to the corresponding content in the above method embodiments, and will not be elaborated here.

[0047] Further, in a preferred embodiment, the DC-powered intelligent lighting control system further includes: The timing control module is configured to perform time period control on the illumination of the target area; The remote control module is configured to remotely adjust and view the illumination state of the target area.

[0048] It can be understood that, in terms of hardware, this embodiment can be implemented by using any existing DC power supply module, intelligent controller, environmental sensor, lighting device and communication module combination, and it can achieve the following effects: Automatic dimming: Detect the ambient brightness through the light sensor, adjust the brightness of the lighting device according to the set threshold, and maintain the stability and comfort of the indoor light.

[0049] Human body induction control: Automatically turn on the illumination when a person is detected to enter, and automatically turn off when leaving, saving energy and providing convenience.

[0050] Timing control: Users can set the timing switch of the lighting system to achieve automatic control by time period, adapting to different usage requirements.

[0051] Remote control: Users can adjust the lighting system through a smartphone APP or other remote control devices, and view and adjust the light state in real time.

[0052] This system aims to provide stable power support for lighting devices through a DC power supply, and at the same time achieve automatic adjustment and high-efficiency management of the lighting system through intelligent control and sensing technologies.

[0053] This embodiment also provides a computer-readable storage medium, on which a direct-current power supply intelligent lighting control program is stored. When the direct-current power supply intelligent lighting control program is executed by a processor, the steps in the above embodiment can be implemented.

[0054] The present invention provides a direct-current power supply intelligent lighting control method and system. It performs human presence detection on a target area to obtain a presence detection result. If the presence detection result indicates that there is a human in the target area, the lighting of the target area is adjusted to a first brightness and maintained at the first brightness state for a preset duration. And it performs a stay behavior detection on the target area to obtain a behavior detection result. If the behavior detection result indicates that the person in the target area has a stay behavior, the lighting of the target area is adjusted to a second brightness, where the second brightness is higher than the first brightness. If the lighting of the target area is in the second brightness state and the presence detection result indicates that there is no human in the target area, the lighting of the target area is turned off. When controlling the lighting, the present invention accurately distinguishes between passing-by and staying scenarios through a two-stage logic of "human presence detection + stay behavior detection", realizes dynamic dimming for maintaining the basic brightness and enhancing the comfortable brightness, solves the defects such as misjudgment and energy consumption waste in the traditional scheme, and solves the problem that the control method of the direct-current power supply lighting system in the prior art is not intelligent enough.

[0055] It should be noted that the various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other.

[0056] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A DC-powered intelligent lighting control method, characterized in that, Including: Performing human presence detection on a target area to obtain a presence detection result; If the presence detection result indicates that there is a human in the target area, adjusting the illumination of the target area to a first brightness and maintaining it for a preset duration in the state of the first brightness; If the presence detection result indicates that there is a human in the target area, performing a stay behavior detection on the target area to obtain a behavior detection result, wherein the operation duration of the stay behavior detection is less than the preset duration; If the behavior detection result indicates that the person in the target area has a stay behavior, adjusting the illumination of the target area to a second brightness, where the second brightness is higher than the first brightness; If the illumination of the target area is in the state of the second brightness and the presence detection result indicates that there is no human in the target area, turning off the illumination of the target area.

2. The DC-powered intelligent lighting control method according to claim 1, wherein Performing human presence detection on a target area to obtain a presence detection result, including: Obtaining a human presence sensing signal, where the human presence sensing signal includes at least one of infrared data, millimeter wave radar data, image data, ultrasonic data, pressure data, or vibration data; Obtaining a presence detection result according to the human presence sensing signal.

3. The DC-powered intelligent lighting control method according to claim 1, wherein Performing a stay behavior detection on a target area to obtain a behavior detection result, including: Obtaining the position state information of the person in the target area; Obtaining a behavior detection result according to the position state information of the person.

4. The DC-powered intelligent lighting control method according to claim 3, wherein Obtaining a behavior detection result according to the position state information of the person, including: Obtaining the environmental information of the target area; Based on a preset neural network model, obtaining a behavior detection result according to the position state information of the person and the environmental information.

5. The DC-powered intelligent lighting control method according to claim 4, wherein Based on a preset neural network model, obtaining a behavior detection result according to the position state information of the person and the environmental information, including: Obtaining the direction angle and velocity component of the person relative to the exit of the target area according to the person state information and the environmental information; Establishing input data according to the direction angle and velocity component; Inputting the input data into the preset neural network model to obtain a behavior detection result.

6. The DC-powered intelligent lighting control method according to claim 5, wherein The preset neural network model includes multiple sub-neural network modules and a comprehensive decision-making module, where the multiple sub-neural network modules correspond one-to-one with multiple exits in the target area and are respectively used to input the input data corresponding to each exit, and the comprehensive decision-making module is connected to the output ends of the multiple sub-neural network modules and is used to output a behavior detection result.

7. The DC-powered intelligent lighting control method according to claim 6, characterized in that The sub-neural network module includes an LSTM layer, and the sub-neural network is used to output a first context vector. The comprehensive decision-making module includes a feedforward neural network layer, and the comprehensive decision-making module is used to input a second context vector formed by weighted summation of multiple first context vectors.

8. A DC-powered intelligent lighting control system, characterized in that, Including: A presence detection module, used to perform human presence detection on a target area to obtain a presence detection result; A first adjustment module, used to adjust the illumination of the target area to a first brightness when the presence detection result indicates that there is a human in the target area and maintain it for a preset duration in the state of the first brightness; A behavior detection module, used to perform a stay behavior detection on the target area to obtain a behavior detection result when the presence detection result indicates that there is a human in the target area, wherein the operation duration of the stay behavior detection is less than the preset duration; The second adjustment module is configured to adjust the illumination of the target area to a second brightness when the behavior detection result indicates that there is a staying behavior of a person in the target area, where the second brightness is higher than the first brightness; The third adjustment module is configured to turn off the illumination of the target area when the illumination of the target area is in the state of the second brightness and the detection result indicates that there is no human body in the target area.

9. The DC-powered intelligent lighting control system according to claim 8, characterized in that, It further includes: The timing control module is configured to perform time period control on the illumination of the target area; The remote control module is configured to remotely adjust and view the illumination state of the target area.

10. A computer-readable storage medium, characterized in that, For storing computer-readable programs or instructions, when the programs or instructions are executed by a processor, the steps in any one of the DC-powered intelligent lighting control methods of claims 1-7 can be implemented.

Images